AI usage and disclosure policy

We use AI systems to help our journalists work faster, never as a substitute for editorial judgement.

What AI helps with

• Monitoring public feeds and official announcements for developing stories.
• Grouping related coverage and identifying which claims need verification.
• Organising source material and preparing first drafts for editors.
• Suggesting headlines, summaries and metadata for editors to accept, edit or reject.

Safeguards

• Human approval by default. No AI-drafted article is published without review and approval by an editor.
• Sourcing requirements. Drafts must cite the sources they rely on. Our systems remove citations that do not match the gathered evidence and flag quotations that cannot be found in source material.
• Extra review for sensitive topics. Security incidents, legal allegations, health or financial claims, politics and personal information are flagged for additional scrutiny.
• Untrusted inputs. Text retrieved from the web is treated as evidence to analyse, never as instructions.
• Disclosure. Articles drafted with substantial AI assistance carry a disclosure note.

Accountability

Editors are accountable for everything we publish, regardless of the tools used to produce it. If you believe an article contains an error, please tell us.